Connecting the Unconnected is a monthly column by the Digital Empowerment Foundation (DEF) that explores how technology can drive inclusion and governance in India. The column focuses on how the digital divide impacts communities differently and advocates for equitable, citizen-informed solutions that ensure technology empowers rather than excludes.
As the India AI Impact Summit 2026 gets underway in Delhi this week, a key focus will be on Digital Public Infrastructure. DPI. What is it? How well can it be conceptualised? And how do we build DPI from the people’s perspective, not from the state’s perspective and certainly not from the point of view of the private sector or corporations?
When India assumed the G20 presidency in 2022, DPI was introduced to ensure ‘no poverty’, ‘zero hunger’, ‘quality education’, ‘good health and well-being’, and ‘affordable clean energy’ – an indication that the Government of India views DPI as a tool to help the poorest of the poor.
Discussions on DPI are often reduced to the systems we have created, such as Aadhaar, UPI, DigiLocker, India Stack, and perhaps BharatNet. But taking a closer look, how many people are actually users of India Stack? The user base is largely top-down.
Even with UPI, we need to ask: who is using it, and who can use it? To use UPI, one needs a smartphone, an app, critical digital literacy, and sufficient bandwidth to complete online transactions. While UPI is the largest ecosystem for financial inclusion, it remains underutilised. If I am digitally illiterate, if I do not have a smartphone, or if I cannot access the internet, I am excluded.
But imagine if the same UPI were available in a village as a shared public infrastructure, a place where people could walk in and do transactions. Then it would truly be a public infrastructure. The question we must ask is: what does people-centric DPI mean, or could mean?
DPI is a public infrastructure, only now it is digital. Public infrastructure must be accessible, like a public good, easily, conveniently, affordably, and for all. A people-centric DPI is one that enables the poorest of the poor – the unconnected, the illiterate, the uneducated, people with disabilities, and marginalised communities – to access services. It should be a rights-based approach like the right to food, the right to education, the right to water, the right to health, the right to information, and access to citizen entitlements. A right without access is meaningless. Likewise, a right without digital meaningful access is futile.
At present, the way DPI is being developed is similar to saying we support zero hunger but without ensuring access. Zero hunger means everyone, especially the poorest, must have access to food. The right to food already exists, but access must be easy, local, affordable, and non-exclusionary. For example, Aadhaar authentication should be easy. A person should be able to walk into a ration shop, collect provisions, and go home. If the authentication infrastructure is located 10 km away, this creates costs, hardship, and exclusion.
India may boast of leading the world in the number of UPI users, with the majority of its population holding digital and biometric IDs such as Aadhaar. But the critical question remains: are these systems accessible at the last mile, near one’s house, within walking distance, as a public good? The answer for a large population is no.
If DPI is like a road, but the road does not reach me, how does it help? Consider artisans: approximately 200 million in rural India, largely unconnected. Or farmers, around 144 million agricultural workers, most of whom are also unconnected. Think of Dalits (200 million), artisans (200 million), tribal communities (104 million), handloom and handicraft workers (3.5 million), panchayat members (3 million), ASHA workers (2 million), and many others.
These populations require a targeted approach to development, and DPI should be designed around their needs. We must ask what kind of systems have we created that exclude 40–50% of people due to limited availability, inaccessibility, unaffordability, and insufficient capacity?
Digital-only policies often emphasise top-down governance rather than bottom-up inclusion. DPI should make life easier for those with the fewest options, not harder.
DPI must be a public sector provision. It should be government-owned and government-accountable. Private-sector support is welcome, but accountability must rest with the government. Like any democratic institution, DPI should be designed for the people, by the people, and with the people.
The voices of the people tell us clearly:
“My attendance is the employer’s responsibility.”
“I pay bribes to middlemen.”
“I walk kilometres for digital access.”
“I cannot read or write.”
These voices must shape DPI design so that grievances do not arise in the first place. This will be the truest people’s DPI, one that aligns with India’s claims at the G20 and serves those who need it most.
People-centric DPI needs to go beyond building platforms and software and ensure that people can use these systems locally, affordably, and without exclusion, through shared access points, assisted services, offline and low-tech options, local languages, and inclusive design. Also, the burden of accessing welfare or proving eligibility must never fall on those with the fewest resources.
People-centric DPI prioritises:
Last-mile reach, so digital services reach the last village and the last household
Ease of use, regardless of literacy, disability, or device ownership
Non-exclusionary access, rejecting digital-only mandates that deny entitlements
Public ownership and data sovereignty, with strong safeguards for privacy and purpose limitation
Bottom-up design, shaped by the voices, needs, and rights of the communities it serves
In essence, people-centric DPI is a digital infrastructure that serves those with no alternatives, ensuring that technology reduces inequality rather than deepening it and that digital governance truly serves public welfare.
Osama Manzar is founder-director of the Digital Empowerment Foundation.
Views expressed are the author’s own.